OpenFox speculative cache warming cuts wait by 10-20s

The technique pre-computes cache entries while the user types, reducing prompt processing time by 10-20 seconds. OpenFox is a free, open-source (MIT) local AI inference harness.